1 重点论文
论文+代码：https://pan.baidu.com/s/1eV-y5bpQIysiurXBD6dylA
1 机械臂视觉抓取
1.1 强化学习+机械臂
1 robopal
Github: https://github.com/NoneJou072/robopal
robopal 是一个基于 MuJoCo 动力学引擎与 pinocchio 机器人动力学库搭建的多平台开源机器人仿真框架，主要用于机械臂的深度强化学习训练与控制算法验证
2 drl_grasping ⭐ 2022 IROS
通过深度强化学习来获得一种稳健的策略，使机器人能够从八叉树形式的紧凑三维观测中抓取各种物体
Github: https://github.com/andrejorsula/drl_grasping
论文: Learning to Grasp on the Moon from 3D Octree Observations with Deep Reinforcement Learning
3 Collaborative-Pushing-Grasping 2022
堆叠物品的分离抓取，用Yolov8进行平面抓取，使用Ubuntu16.04
Github: https://github.com/nizhihao/Collaborative-Pushing-Grasping
论文: Collaborative Pushing and Grasping of Tightly Stacked Objects via Deep Reinforcement Learning
4 Eureka 2023
通过编写大型语言模型进行人性化的奖励设计，Github很多star，LLM+RL
Github: https://github.com/eureka-research/Eureka
论文: Eureka Human-Level Reward Design via Coding Large Language Models
1.2 大语言模型+机械臂
1 VoxPoser 2023 ArXiv
零样本操控机器人，李飞飞团队新作Voxposer
Project page: https://voxposer.github.io/ (代码暂未开源)
论文: VoxPoser Composable 3D Value Maps for Robotic Manipulation with Language Models for Robotic Manipulation with Language Models
2 Language to Rewards for Robotic Skill Synthesis 2023 CoRL
利用大型语言模型 (LLM) 和奖励函数来优化机器人控制的方法
Project page: https://language-to-reward.github.io/ (代码暂未开源)
论文: Language to Rewards for Robotic Skill Synthesis
3 RobotCat 2023 ArXiv
谷歌DeepMind推出了一种可以自我改进、自我提升（self-improving）的用于机器人的AI智能体
Github: https://github.com/kyegomez/RoboCAT
论文: RoboCat A Self-Improving Foundation Agent for Robotic Manipulation
4 RoboAgent 2023 ArXiv
一种AI智能体，利用被动观察和主动学习，使机器人能够获得与幼儿同等的操作能力
Project page: https://robopen.github.io/
论文: RoboAgent Generalization and Efficiency in Robot Manipulation via Semantic Augmentations and Action Chunking
1.3 其他视觉抓取
1 GraspNeRF 2023 ICRA
使用通用 NeRF 对透明和镜面物体进行基于多视图的 6-DoF 抓取检测
Project page: https://pku-epic.github.io/GraspNeRF/
论文: GraspNeRF Multiview-based 6-DoF Grasp Detection for Transparent and Specular Objects Using Generalizable NeRF
2 Transporter Networks 2020 CoRL
基于视觉操作实现重新整理目标物体
Project page: https://transporternets.github.io/
论文: Transporter Networks Rearranging the Visual World for Robotic Manipulatio
3 CLIPort 2021 CoRL
一个端到端模仿学习agent，可以为各种桌面任务学习单一语言条件策略
Project page: https://cliport.github.io/
论文: CLIPORT What and Where Pathways for Robotic Manipulation
4 CaTGrasp 2022 ICRA⭐
任务驱动的类级别机器人抓取
Github: https://github.com/wenbowen123/catgrasp
论文: CaTGrasp_Learning_Category-Level_Task-Relevant_Grasping_in_Clutter_from_Simulation